Refactor: Always include risk fields #1052

malhotra5 · 2025-11-06T15:49:07Z

Summary

This PR does the following

makes sure tool schemas is always has the security_risk field
validates security_risk field at runtime.
- when llm security analyzer is configured + security_risk is missing -> send an error event to correct itself (good for strong models)
- when llm security analyzer is NOT configured + security_risk is missing -> default to SecurityRisk.UNKNOWN and proceed as usually
move security analyzer to conversation state to support easier mid conversation switches

Benefits

Consistent schemas
Runtime enforcement of security_risk provides the desired flexibility
Graceful error handling (we send back an error event so the model has a chance to retry, good for stronger models)
No compromise on traceability (we have events registering when the security analyzer is added to a conversation; this is important since every event will contain security_risk field and they can be ignored depending on whether the analyzer has been configured)

The linked issue contains the following scenarios; this PR helps address each one of them

Scenario 1:
Security analyzer exists when starting conversation
System prompt includes tools with security_risk
Security analyzer is removed mid conversation
No risk field is expected, but LLM still believes it needs to pass this field because system prompt was not updated. Passing field will crash the conversation

In this scenario, the security_risk parameter will still exist. The LLM can continue passing the field. We simply won't interfere with the control loop even though the agent is passing risk values. The agent can also forget to pass a risk parameter which will be ignored.

Scenario 2:
Security analyzer was not included at start of conversation
System prompt does not include information on passing security_risk fields
Security analyzer is added later in the conversation
security_risk is a required field, but LLM is unaware of this requirement since system prompt did not include this information

This time security_risk information in the system prompt will always exist at the start of a conversation. The LLM is always aware of it. If analyzer is not configured and the LLM forgets to pass a value, we'll default to SecurityRisk.UNKNOWN and proceed as usual. When security analyzer is added later in the conversation, we'll make sure that at runtime an error event is raised if the LLM fails to pass the risk parameter.

Scenario 3:
Similar to 1, but:
The security analyzer is removed mid conversation
The conversation is fully reloaded
ActionEvents are created in history, and they have the old security_risk field
Validation fails, and leads to an infinite loop

The schema should always contain security_risk, so loading back events regardless of whether a section of conversation had or did not have a security analyzer should work.

Fixes #819

@malhotra5 can click here to continue refining the PR

Agent Server images for this PR

• GHCR package: https://github.com/OpenHands/agent-sdk/pkgs/container/agent-server

Variants & Base Images

Variant	Architectures	Base Image	Docs / Tags
java	amd64, arm64	`eclipse-temurin:17-jdk`	Link
python	amd64, arm64	`nikolaik/python-nodejs:python3.12-nodejs22`	Link
golang	amd64, arm64	`golang:1.21-bookworm`	Link

Pull (multi-arch manifest)

# Each variant is a multi-arch manifest supporting both amd64 and arm64
docker pull ghcr.io/openhands/agent-server:8aaca4b-python

Run

docker run -it --rm \
  -p 8000:8000 \
  --name agent-server-8aaca4b-python \
  ghcr.io/openhands/agent-server:8aaca4b-python

All tags pushed for this build

ghcr.io/openhands/agent-server:8aaca4b-golang-amd64
ghcr.io/openhands/agent-server:8aaca4b-golang_tag_1.21-bookworm-amd64
ghcr.io/openhands/agent-server:8aaca4b-golang-arm64
ghcr.io/openhands/agent-server:8aaca4b-golang_tag_1.21-bookworm-arm64
ghcr.io/openhands/agent-server:8aaca4b-java-amd64
ghcr.io/openhands/agent-server:8aaca4b-eclipse-temurin_tag_17-jdk-amd64
ghcr.io/openhands/agent-server:8aaca4b-java-arm64
ghcr.io/openhands/agent-server:8aaca4b-eclipse-temurin_tag_17-jdk-arm64
ghcr.io/openhands/agent-server:8aaca4b-python-amd64
ghcr.io/openhands/agent-server:8aaca4b-nikolaik_s_python-nodejs_tag_python3.12-nodejs22-amd64
ghcr.io/openhands/agent-server:8aaca4b-python-arm64
ghcr.io/openhands/agent-server:8aaca4b-nikolaik_s_python-nodejs_tag_python3.12-nodejs22-arm64
ghcr.io/openhands/agent-server:8aaca4b-golang
ghcr.io/openhands/agent-server:8aaca4b-java
ghcr.io/openhands/agent-server:8aaca4b-python

About Multi-Architecture Support

Each variant tag (e.g., 8aaca4b-python) is a multi-arch manifest supporting both amd64 and arm64
Docker automatically pulls the correct architecture for your platform
Individual architecture tags (e.g., 8aaca4b-python-amd64) are also available if needed

- Modified tool schema generation to include security_risk field when add_security_risk_prediction=True for all tool types (including read-only tools) - Updated LLM security analyzer validation to always require security_risk field when using LLMSecurityAnalyzer - Added comprehensive test suite for security_risk validation behavior - Fixed existing tests to reflect new behavior where security_risk is included for read-only tools when prediction is enabled - Updated docstrings to clarify the new behavior Co-authored-by: openhands <openhands@all-hands.dev>

github-actions · 2025-11-06T15:52:20Z

Coverage Report •

File	Stmts	Miss	Cover	Missing
openhands-agent-server/openhands/agent_server
conversation_router.py	97	48	50%	80–82, 96–97, 110, 121–123, 151–154, 165–168, 185, 189–191, 197, 211–213, 215, 217, 219–221, 234–238, 251–255, 270–273, 286, 289–291
event_service.py	204	86	57%	52–53, 72–74, 77–82, 94, 110, 117, 119, 126–127, 135–138, 145–147, 159–160, 163–164, 166–168, 170, 175, 178, 182–184, 186, 188, 192–196, 200, 204–205, 207, 224–225, 272, 279–280, 282, 285–287, 289, 293–296, 300–303, 311–314, 333–334, 336–343, 345–346, 352, 358, 371–372
models.py	98	2	97%	56–57
openhands-sdk/openhands/sdk/agent
agent.py	167	55	67%	80, 84, 141, 145–146, 153–154, 156–158, 160–162, 178, 193–195, 202–204, 206, 210, 213–214, 216, 223, 250, 255, 275, 282, 304, 309, 320, 323, 346, 356–357, 378–380, 382, 394–395, 400–401, 421–422, 427, 439–440, 445–446, 478, 485–486
base.py	177	24	86%	163, 181, 187, 199–201, 227–228, 240–241, 247–249, 262, 270–271, 302, 346, 353, 366, 403–404, 414–415
openhands-sdk/openhands/sdk/conversation
base.py	69	5	92%	106, 111–112, 147, 158
state.py	132	12	90%	175, 200, 247–249, 265–266, 295, 313, 322, 337, 343
openhands-sdk/openhands/sdk/conversation/impl
local_conversation.py	189	53	71%	121, 123–124, 157–158, 181, 186, 198, 215, 223–225, 229–230, 282–283, 286, 293, 314–316, 319, 328, 344, 346, 348, 352, 354–356, 358, 360, 366–367, 380–381, 383, 385, 389–392, 409–410, 415, 420, 422, 427, 429–431, 449, 451
remote_conversation.py	372	104	72%	55–61, 68–71, 100, 107, 115, 117–120, 130, 139, 143–144, 149–152, 187, 201, 218, 229, 238–239, 291, 311, 319, 331, 339–342, 345, 350–353, 355, 360–361, 366–370, 375–379, 384–387, 390, 401–402, 406, 410, 413, 484, 490, 492, 508, 510–511, 522, 539–540, 546, 561, 594, 596, 598–599, 603–604, 613–614, 623, 631, 636–638, 640, 643, 645–646, 663, 670, 676–677, 691–692, 699–700
TOTAL	12046	5552	53%

…ation - Created new SecurityAnalyzerConfigurationEvent class extending Event - Added event type to EventType literal and exports - Modified AgentBase.init_state to always emit SecurityAnalyzerConfigurationEvent - Added comprehensive tests for event creation and emission - Event tracks analyzer type (string name or None) and includes visualization methods Co-authored-by: openhands <openhands@all-hands.dev>

- Test all 5 required scenarios with parameterized tests - Case 1: LLM analyzer set, security risk passed, extracted properly - Case 2: analyzer not set, security risk passed, extracted properly - Case 3: LLM analyzer set, security risk not passed, ValueError raised - Case 4: analyzer not set, security risk not passed, UNKNOWN returned - Case 5: invalid security risk value passed, ValueError raised - Include additional tests for error messages and argument mutation - Follow existing test patterns and code style guidelines Co-authored-by: openhands <openhands@all-hands.dev>

…ehavior in conversations - Test new conversation initialization creates SystemPromptEvent and SecurityAnalyzerConfigurationEvent - Test reinitializing with same analyzer type creates new events (different instances) - Test reinitializing with same agent instance still creates new events - Test switching between different analyzers creates appropriate events - Test switching from no analyzer to analyzer creates events - Test multiple reinitializations create correct event sequences - Test event properties and methods validation - Use parameterized tests and fixtures following existing patterns - All 8 tests passing with proper edge case coverage Co-authored-by: openhands <openhands@all-hands.dev>

- Updated test_conversation_event_id_validation to expect duplicate event at index 2 instead of 1 - Updated test_pause_basic_functionality to expect 2 events instead of 1 - Both tests now account for the new SecurityAnalyzerConfigurationEvent being added during conversation initialization Co-authored-by: openhands <openhands@all-hands.dev>

…gurationEvent - Updated expected event count to account for additional SecurityAnalyzerConfigurationEvent - When conversation is loaded from persistence, agent initialization adds one more event - Changed assertion from original_event_count to original_event_count + 1 Co-authored-by: openhands <openhands@all-hands.dev>

- Added __eq__ method to SecurityAnalyzerConfigurationEvent to compare only analyzer_type - This prevents duplicate events when the same analyzer configuration is used - Updated test expectations to account for SecurityAnalyzerConfigurationEvent being added during initialization - Fixed test_conversation_event_id_validation to expect duplicate at index 2 (after SystemPromptEvent and SecurityAnalyzerConfigurationEvent) - Fixed test_conversation_persistence_lifecycle to expect same event count when loading from persistence Co-authored-by: openhands <openhands@all-hands.dev>

…//github.com/OpenHands/software-agent-sdk into refactor/always-include-security-risk-fields

This reverts commit 160a6a2.

- Fix _extract_security_risk method to always pop security_risk from arguments before checking readOnlyHint - Update test calls to include the new readOnlyHint parameter (3rd argument) - Add comprehensive test for readOnlyHint=True scenario - Ensure security_risk is properly removed from arguments in all cases Co-authored-by: openhands <openhands@all-hands.dev>

malhotra5 · 2025-11-11T16:26:49Z

QA'ed with CLI and seems to work well

openhands-sdk/openhands/sdk/conversation/state.py

openhands-sdk/openhands/sdk/agent/agent.py

xingyaoww · 2025-11-11T17:16:30Z

openhands-sdk/openhands/sdk/agent/base.py

+        if "security_analyzer" in d and d["security_analyzer"]:
+            warnings.warn(
+                AGENT_SECURITY_ANALYZER_DEPRECATION_WARNING,
+                DeprecationWarning,
+                stacklevel=3,
+            )


@OpenHands is there any good python library that allow us to manage deprecation warning in a more systematic way?

I'm on it! xingyaoww can track my progress at all-hands.dev

Summary

I answered your question about Python libraries for managing deprecation warnings systematically. Here are the recommendations:

Top 3 Recommended Libraries:

deprecation (Best for your use case)

Lightweight and purpose-built for deprecation management

Supports version tracking (deprecated_in, removed_in, current_version)

Integrates seamlessly with Python's warnings module

Ideal for SDKs and libraries

Installation: pip install deprecation

Deprecated (deprecated.py)

Feature-rich with excellent Sphinx documentation integration

Simple decorator syntax with version and reason tracking

Very popular (4M+ downloads/month)

Installation: pip install Deprecated

debtcollector (From OpenStack)

Comprehensive solution for tracking both deprecations and technical debt

Advanced features like move notifications and removal tracking

Battle-tested in large projects

Installation: pip install debtcollector

Why deprecation is recommended for your codebase:

Looking at your current usage (multiple warnings.warn calls in agent/base.py, llm/llm.py, and conversation/conversation_stats.py), the deprecation library would provide:

Version tracking for deprecation timelines

Minimal overhead and dependencies

Easy migration from your current warnings usage

Clear, structured messaging

SDK-friendly implementation

All three libraries provide decorators that make deprecation management more systematic and maintainable than the built-in warnings module alone.

No code changes were made as this was an informational question.

_{View full conversation}

xingyaoww · 2025-11-11T17:42:40Z

tests/agent_server/test_conversation_router.py

@OpenHands please add a test to tests/cross/test_remote_conversation_live_server.py where we spawns a real agent server to test this

I'm on it! xingyaoww can track my progress at all-hands.dev

Summary

I've successfully added an integration test to tests/cross/test_remote_conversation_live_server.py that spawns a real agent server to test the security_risk field functionality introduced in PR #1052.

Changes Made

Added test: test_security_risk_field_with_live_server (~148 lines)

Spawns a real agent server via the existing server_env fixture

Uses a monkeypatched LLM completion that returns tool calls without the security_risk field

Validates that:

Actions execute successfully even when security_risk is missing (defaults to UNKNOWN)

ActionEvent objects always have a security_risk attribute

Test Details

The test:

Creates an agent with LLMSecurityAnalyzer configured

Mocks LLM completions to return a finish tool call without providing the security_risk parameter

Verifies the action executes successfully (demonstrating the graceful default behavior)

Confirms that the resulting ActionEvent has the security_risk attribute

Validation

✅ Test passes successfully
✅ Pre-commit checks pass (ruff format, ruff lint, pycodestyle, pyright)
✅ Changes committed and pushed to the refactor/always-include-security-risk-fields branch

This integration test provides end-to-end validation of the security_risk field handling changes, ensuring the system gracefully handles cases where the field is missing while still maintaining backward compatibility.

_{View full conversation}

github-actions · 2025-11-11T17:42:56Z

Hi! I started running the integration tests on your PR. You will receive a comment with the results shortly.

xingyaoww

LGTM! Will merge this after tests pass

github-actions · 2025-11-11T17:45:59Z

🔄 Running Examples with `openhands/claude-haiku-4-5-20251001`

Last updated: 2025-11-11 18:07:13 UTC

Example	Status	Duration	Cost
01_standalone_sdk/02_custom_tools.py	✅ PASS	27s	$0.03
01_standalone_sdk/03_activate_skill.py	✅ PASS	12s	$0.01
01_standalone_sdk/05_use_llm_registry.py	✅ PASS	10s	$0.01
01_standalone_sdk/07_mcp_integration.py	❌ FAIL (exit: 1)	35s	$0.00
01_standalone_sdk/09_pause_example.py	✅ PASS	13s	$0.01
01_standalone_sdk/10_persistence.py	✅ PASS	38s	$0.02
01_standalone_sdk/11_async.py	✅ PASS	42s	$0.03
01_standalone_sdk/12_custom_secrets.py	✅ PASS	14s	$0.01
01_standalone_sdk/13_get_llm_metrics.py	✅ PASS	33s	$0.02
01_standalone_sdk/14_context_condenser.py	✅ PASS	202s	$0.35
01_standalone_sdk/17_image_input.py	✅ PASS	21s	$0.02
01_standalone_sdk/18_send_message_while_processing.py	✅ PASS	19s	$0.01
01_standalone_sdk/19_llm_routing.py	✅ PASS	20s	$0.02
01_standalone_sdk/20_stuck_detector.py	✅ PASS	19s	$0.02
01_standalone_sdk/21_generate_extraneous_conversation_costs.py	✅ PASS	17s	$0.01
01_standalone_sdk/22_anthropic_thinking.py	✅ PASS	15s	$0.01
01_standalone_sdk/23_responses_reasoning.py	✅ PASS	46s	$0.01
01_standalone_sdk/24_planning_agent_workflow.py	✅ PASS	221s	$0.23
01_standalone_sdk/25_agent_delegation.py	❌ FAIL (exit: 1)	67s	$0.00
01_standalone_sdk/26_custom_visualizer.py	✅ PASS	24s	$0.00N/A
02_remote_agent_server/01_convo_with_local_agent_server.py	✅ PASS	69s	$0.06
02_remote_agent_server/02_convo_with_docker_sandboxed_server.py	✅ PASS	136s	$0.07
02_remote_agent_server/03_browser_use_with_docker_sandboxed_server.py	✅ PASS	67s	$0.05
02_remote_agent_server/04_convo_with_api_sandboxed_server.py	✅ PASS	90s	$0.02

❌ Some tests failed

Total: 24 | Passed: 22 | Failed: 2

View full workflow run

github-actions · 2025-11-11T17:47:40Z

🧪 Integration Tests Results

Overall Success Rate: 93.8%
Total Cost: $0.69
Models Tested: 4
Timestamp: 2025-11-11 17:47:38 UTC

📁 Detailed Logs & Artifacts

Click the links below to access detailed agent/LLM logs showing the complete reasoning process for each model. On the GitHub Actions page, scroll down to the 'Artifacts' section to download the logs.

litellm_proxy_gpt_5_mini_2025_08_07: 📥 View & Download Logs
litellm_proxy_claude_sonnet_4_5_20250929: 📥 View & Download Logs
litellm_proxy_deepseek_deepseek_chat: 📥 View & Download Logs
litellm_proxy_moonshot_kimi_k2_thinking: 📥 View & Download Logs

📊 Summary

Model	Success Rate	Tests Passed	Skipped	Total Tests	Cost
litellm_proxy_gpt_5_mini_2025_08_07	100.0%	8/8	0	8	$0.05
litellm_proxy_claude_sonnet_4_5_20250929	100.0%	8/8	0	8	$0.41
litellm_proxy_deepseek_deepseek_chat	87.5%	7/8	1	8	$0.03
litellm_proxy_moonshot_kimi_k2_thinking	87.5%	7/8	1	8	$0.20

📋 Detailed Results

litellm_proxy_gpt_5_mini_2025_08_07

Success Rate: 100.0% (8/8)
Total Cost: $0.05
Run Suffix: litellm_proxy_gpt_5_mini_2025_08_07_41101cc_gpt5_mini_run_N8_20251111_174328

litellm_proxy_claude_sonnet_4_5_20250929

Success Rate: 100.0% (8/8)
Total Cost: $0.41
Run Suffix: litellm_proxy_claude_sonnet_4_5_20250929_41101cc_sonnet_run_N8_20251111_174323

litellm_proxy_deepseek_deepseek_chat

Success Rate: 87.5% (7/8)
Total Cost: $0.03
Run Suffix: litellm_proxy_deepseek_deepseek_chat_41101cc_deepseek_run_N8_20251111_174318
Skipped Tests: 1

Skipped Tests:

t08_image_file_viewing: This test requires a vision-capable LLM model. Please use a model that supports image input.

litellm_proxy_moonshot_kimi_k2_thinking

Success Rate: 87.5% (7/8)
Total Cost: $0.20
Run Suffix: litellm_proxy_moonshot_kimi_k2_thinking_41101cc_kimi_k2_run_N8_20251111_174322
Skipped Tests: 1

Skipped Tests:

t08_image_file_viewing: This test requires a vision-capable LLM model. Please use a model that supports image input.

This test validates the fix for issue #819 where security_risk field handling was inconsistent. It spawns a real agent server and tests that: 1. Actions can be executed without security_risk (defaults to UNKNOWN) 2. ActionEvent always has a security_risk attribute The test uses a monkeypatched LLM completion to return a finish tool call without the security_risk field, ensuring the system handles it gracefully. Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww

LGTM!

malhotra5 and others added 26 commits November 6, 2025 10:53

Update agent.py

0e5b697

send back error events

e087266

simplify risk field handling

7b5a9dd

fix comment

c5e6329

add comments

90eeb48

move event emit

42a2bfc

prevent dupe configuration events

e1e612c

rm fluff tests

951a405

write tests for event equality and serialization

35ce7db

Merge branch 'main' into refactor/always-include-security-risk-fields

625e509

fix merge conflicts

1106bdf

fix tests

e188cf3

Merge branch 'main' into refactor/always-include-security-risk-fields

480e53d

always default to adding risk prediction

160a6a2

Merge branch 'refactor/always-include-security-risk-fields' of https:…

8405598

…//github.com/OpenHands/software-agent-sdk into refactor/always-include-security-risk-fields

Revert "always default to adding risk prediction"

01646be

This reverts commit 160a6a2.

Update tool.py

248648f

simplify

2402da2

handle readonly case

4b3e817

malhotra5 force-pushed the refactor/always-include-security-risk-fields branch from 3a8f6bb to 4d58824 Compare November 7, 2025 16:09

Delete test_security_risk_schema_consistency.py

e222455

malhotra5 added 4 commits November 11, 2025 11:10

Merge branch 'main' into refactor/always-include-security-risk-fields

533f7b0

rm old event

f314857

Update types.py

090ae58

simplify tests

eca4672

malhotra5 marked this pull request as ready for review November 11, 2025 16:26

xingyaoww reviewed Nov 11, 2025

View reviewed changes

openhands-sdk/openhands/sdk/conversation/state.py Outdated Show resolved Hide resolved

openhands-sdk/openhands/sdk/agent/agent.py Show resolved Hide resolved

malhotra5 marked this pull request as draft November 11, 2025 17:02

malhotra5 added 2 commits November 11, 2025 12:04

rm configuration history

f973187

document decision via comment

b5774dd

malhotra5 marked this pull request as ready for review November 11, 2025 17:14

xingyaoww reviewed Nov 11, 2025

View reviewed changes

use model validator for security prompt default

d60da37

xingyaoww mentioned this pull request Nov 11, 2025

Adopt a systematic deprecation management library #1137

Open

Merge branch 'main' into refactor/always-include-security-risk-fields

41101cc

xingyaoww reviewed Nov 11, 2025

View reviewed changes

xingyaoww added the integration-test Runs the integration tests and comments the results label Nov 11, 2025

xingyaoww reviewed Nov 11, 2025

View reviewed changes

xingyaoww added the test-examples Run all applicable "examples/" files. Expensive operation. label Nov 11, 2025

openhands-agent and others added 2 commits November 11, 2025 17:58

Merge branch 'main' into refactor/always-include-security-risk-fields

7b7a325

xingyaoww enabled auto-merge (squash) November 11, 2025 18:00

xingyaoww disabled auto-merge November 11, 2025 18:00

xingyaoww approved these changes Nov 11, 2025

View reviewed changes

xingyaoww enabled auto-merge (squash) November 11, 2025 18:01

xingyaoww merged commit a656ca7 into main Nov 11, 2025
15 checks passed

xingyaoww deleted the refactor/always-include-security-risk-fields branch November 11, 2025 18:02

Refactor: Always include risk fields #1052

Refactor: Always include risk fields #1052

Uh oh!

Conversation

malhotra5 commented Nov 6, 2025 • edited by github-actions bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Uh oh!

github-actions bot commented Nov 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

malhotra5 commented Nov 11, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

xingyaoww Nov 11, 2025

Choose a reason for hiding this comment

Uh oh!

openhands-ai bot Nov 11, 2025

Choose a reason for hiding this comment

Uh oh!

openhands-ai bot Nov 11, 2025

Choose a reason for hiding this comment

Summary

Top 3 Recommended Libraries:

Why deprecation is recommended for your codebase:

Uh oh!

xingyaoww Nov 11, 2025

Choose a reason for hiding this comment

Uh oh!

openhands-ai bot Nov 11, 2025

Choose a reason for hiding this comment

Uh oh!

openhands-ai bot Nov 11, 2025

Choose a reason for hiding this comment

Summary

Changes Made

Test Details

Validation

Uh oh!

github-actions bot commented Nov 11, 2025

Uh oh!

xingyaoww left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Nov 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔄 Running Examples with openhands/claude-haiku-4-5-20251001

❌ Some tests failed

Uh oh!

github-actions bot commented Nov 11, 2025

🧪 Integration Tests Results

📁 Detailed Logs & Artifacts

📊 Summary

📋 Detailed Results

litellm_proxy_gpt_5_mini_2025_08_07

litellm_proxy_claude_sonnet_4_5_20250929

litellm_proxy_deepseek_deepseek_chat

litellm_proxy_moonshot_kimi_k2_thinking

Uh oh!

xingyaoww left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

malhotra5 commented Nov 6, 2025 •

edited by github-actions bot

Loading

github-actions bot commented Nov 6, 2025 •

edited

Loading

Why `deprecation` is recommended for your codebase:

github-actions bot commented Nov 11, 2025 •

edited

Loading

🔄 Running Examples with `openhands/claude-haiku-4-5-20251001`